The aim of Tajfel's research was to investigate if intergroup discrimination would take place based on being put into different groups. Tajfel hypothesized that categorization and discrimination operate automatically, even when there is not necessarily any prior prejudice.

A sample of 48 boys, ages 14 - 15 years old, were asked to rate 12 paintings by the abstract expressionist painters Paul Klee and Wassily Kandinsky. They were not aware of which artist had painted which painting during the test. The boys were then randomly allocated to one of two groups and told that they had preferred either Klee or Kandinsky.

Each boy was then given the task to award points to two other boys, one from his same group and one from the other group. The only information that each boy was given was the code numbers and the name of the group of the two boys they were supposed to award. There were two systems of awarding points that were employed by the researchers.

Point allocation system 1: The point scores for each boy were linked so that the sum of the two scores was 15.  If a participant chose 8, the other boy automatically got a score of 7 (15 - 8). This means that as the score for the participant increases, the score of the other boy decreases.

Point allocation system 2: In the second method of allocating points, Tajfel manipulated the system.

• If a Klee member chose a high value for another Klee member, it would give a higher profit to the out-group.
• If a Klee member chose a mid-range value for another Klee member, it would give the same points for the other group.
• If a Klee member chose a low value for another Klee member, it would award only 1 point to the other team.

In the first system of point allocation, the boys generally awarded more points to the members of their in-group showing in-group favouritism. In the second system of point awarding, the boys were willing to give their own team fewer points with the goal of maximizing the difference between their in-group and the out-group.  This was a bit surprising since it meant that the boys left the study with fewer points than if they had all given each other the largest number of points possible.

One of the most obvious conclusions that we can draw from this experiment is the natural tendency of members of a group to favour their in-group. Despite the seemingly meaningless groupings created by the experimenters, the participants were able to identify with their respective groups and create a positive social identity by giving their in-group more points. Tajfel demonstrated that a "minimal group" is all that is necessary for individuals to exhibit discrimination against an out-group. This experiment is considered a classic in psychology because it demonstrates that intergroup conflict is not required for discrimination to occur. The study thus challenged previous beliefs that competition was necessary and sufficient to produce prejudice.

A follow-up study done by Tajfel & Billig (1973) showed that even when members of the groups were aware that the groupings were completely random and not based on any criteria, the participants still showed in-group favouritism.

• The experiment had a high level of control.  Confounding variables were minimized.
• The task the participants were asked to was highly artificial; the study lacks ecological validity.  This may not reflect actual behaviour in a naturalistic setting.
• The boys may have shown demand characteristics, trying to please the researcher.  The boys may have interpreted the task as competitive and tried to 'win."
• The procedure can be replicated to establish reliability.
• Sampling bias - the study was carried out on British schoolboys. It is difficult to generalize the results to women, adults or other cultures.